1 paper
Chengzhi Wu, Xuelei Bi, Julius Pfrommer +3
On robotics computer vision tasks, generating and annotating large amounts of data from real-world for the use of deep learning-based approaches is often difficult or even impossib…